includes/filerepo/backend/FileBackend.php

   1 <?php
   2 /**
   3  * @defgroup FileBackend File backend
   4  * @ingroup  FileRepo
   5  *
   6  * File backend is used to interact with file management systems such as SWIFT.
   7  *
   8  */
   9
  10 /**
  11  * @file
  12  * @ingroup FileBackend
  13  * @author Aaron Schulz
  14  */
  15
  16 /**
  17  * Base class for all file backend classes (including multi-write backends).
  18  *
  19  * This class defines the methods as abstract that subclasses must implement.
  20  * Outside callers can assume that all backends will have these functions.
  21  *
  22  * All "storage paths" are of the format "mwstore://backend/container/path".
  23  * The paths use UNIX file system (FS) notation, though any particular backend may
  24  * not actually be using a local filesystem. Therefore, the paths are only virtual.
  25  *
  26  * Backend contents are stored under wiki-specific container names by default.
  27  * For legacy reasons, this has no effect for the FS backend class, and per-wiki
  28  * segregation must be done by setting the container paths appropriately.
  29  *
  30  * FS-based backends are somewhat more restrictive due to the existence of real
  31  * directory files; a regular file cannot have the same name as a directory. Other
  32  * backends with virtual directories may not have this limitation. Callers should
  33  * store files in such a way that no files and directories are under the same path.
  34  *
  35  * Methods should avoid throwing exceptions at all costs.
  36  * As a corollary, external dependencies should be kept to a minimum.
  37  *
  38  * @ingroup FileBackend
  39  * @since 1.19
  40  */
  41 abstract class FileBackend {
  42         protected $name; // string; unique backend name
  43         protected $wikiId; // string; unique wiki name
  44         protected $readOnly; // string; read-only explanation message
  45         /** @var LockManager */
  46         protected $lockManager;
  47
  48         /**
  49          * Create a new backend instance from configuration.
  50          * This should only be called from within FileBackendGroup.
  51          *
  52          * $config includes:
  53          *     'name'        : The unique name of this backend.
  54          *                     This should consist of alphanumberic, '-', and '_' characters.
  55          *     'wikiId'      : Prefix to container names that is unique to this wiki.
  56          *                     This should consist of alphanumberic, '-', and '_' characters.
  57          *     'lockManager' : Registered name of a file lock manager to use.
  58          *     'readOnly'    : Write operations are disallowed if this is a non-empty string.
  59          *                     It should be an explanation for the backend being read-only.
  60          *
  61          * @param $config Array
  62          */
  63         public function __construct( array $config ) {
  64                 $this->name = $config['name'];
  65                 $this->wikiId = isset( $config['wikiId'] )
  66                         ? $config['wikiId']
  67                         : wfWikiID(); // e.g. "my_wiki-en_"
  68                 $this->lockManager = ( $config['lockManager'] instanceof LockManager )
  69                         ? $config['lockManager']
  70                         : LockManagerGroup::singleton()->get( $config['lockManager'] );
  71                 $this->readOnly = isset( $config['readOnly'] )
  72                         ? (string)$config['readOnly']
  73                         : '';
  74         }
  75
  76         /**
  77          * Get the unique backend name.
  78          * We may have multiple different backends of the same type.
  79          * For example, we can have two Swift backends using different proxies.
  80          *
  81          * @return string
  82          */
  83         final public function getName() {
  84                 return $this->name;
  85         }
  86
  87         /**
  88          * Check if this backend is read-only
  89          *
  90          * @return bool
  91          */
  92         final public function isReadOnly() {
  93                 return ( $this->readOnly != '' );
  94         }
  95
  96         /**
  97          * Get an explanatory message if this backend is read-only
  98          *
  99          * @return string|false Returns falls if the backend is not read-only
 100          */
 101         final public function getReadOnlyReason() {
 102                 return ( $this->readOnly != '' ) ? $this->readOnly : false;
 103         }
 104
 105         /**
 106          * This is the main entry point into the backend for write operations.
 107          * Callers supply an ordered list of operations to perform as a transaction.
 108          * Files will be locked, the stat cache cleared, and then the operations attempted.
 109          * If any serious errors occur, all attempted operations will be rolled back.
 110          *
 111          * $ops is an array of arrays. The outer array holds a list of operations.
 112          * Each inner array is a set of key value pairs that specify an operation.
 113          *
 114          * Supported operations and their parameters:
 115          * a) Create a new file in storage with the contents of a string
 116          *     array(
 117          *         'op'                  => 'create',
 118          *         'dst'                 => <storage path>,
 119          *         'content'             => <string of new file contents>,
 120          *         'overwrite'           => <boolean>,
 121          *         'overwriteSame'       => <boolean>
 122          *     )
 123          * b) Copy a file system file into storage
 124          *     array(
 125          *         'op'                  => 'store',
 126          *         'src'                 => <file system path>,
 127          *         'dst'                 => <storage path>,
 128          *         'overwrite'           => <boolean>,
 129          *         'overwriteSame'       => <boolean>
 130          *     )
 131          * c) Copy a file within storage
 132          *     array(
 133          *         'op'                  => 'copy',
 134          *         'src'                 => <storage path>,
 135          *         'dst'                 => <storage path>,
 136          *         'overwrite'           => <boolean>,
 137          *         'overwriteSame'       => <boolean>
 138          *     )
 139          * d) Move a file within storage
 140          *     array(
 141          *         'op'                  => 'move',
 142          *         'src'                 => <storage path>,
 143          *         'dst'                 => <storage path>,
 144          *         'overwrite'           => <boolean>,
 145          *         'overwriteSame'       => <boolean>
 146          *     )
 147          * e) Delete a file within storage
 148          *     array(
 149          *         'op'                  => 'delete',
 150          *         'src'                 => <storage path>,
 151          *         'ignoreMissingSource' => <boolean>
 152          *     )
 153          * f) Do nothing (no-op)
 154          *     array(
 155          *         'op'                  => 'null',
 156          *     )
 157          *
 158          * Boolean flags for operations (operation-specific):
 159          * 'ignoreMissingSource' : The operation will simply succeed and do
 160          *                         nothing if the source file does not exist.
 161          * 'overwrite'           : Any destination file will be overwritten.
 162          * 'overwriteSame'       : An error will not be given if a file already
 163          *                         exists at the destination that has the same
 164          *                         contents as the new contents to be written there.
 165          *
 166          * $opts is an associative of boolean flags, including:
 167          * 'force'               : Errors that would normally cause a rollback do not.
 168          *                         The remaining operations are still attempted if any fail.
 169          * 'nonLocking'          : No locks are acquired for the operations.
 170          *                         This can increase performance for non-critical writes.
 171          *                         This has no effect unless the 'force' flag is set.
 172          * 'allowStale'          : Don't require the latest available data.
 173          *                         This can increase performance for non-critical writes.
 174          *                         This has no effect unless the 'force' flag is set.
 175          *
 176          * Remarks on locking:
 177          * File system paths given to operations should refer to files that are
 178          * already locked or otherwise safe from modification from other processes.
 179          * Normally these files will be new temp files, which should be adequate.
 180          *
 181          * Return value:
 182          * This returns a Status, which contains all warnings and fatals that occured
 183          * during the operation. The 'failCount', 'successCount', and 'success' members
 184          * will reflect each operation attempted. The status will be "OK" unless any
 185          * of the operations failed and the 'force' parameter was not set.
 186          *
 187          * @param $ops Array List of operations to execute in order
 188          * @param $opts Array Batch operation options
 189          * @return Status
 190          */
 191         final public function doOperations( array $ops, array $opts = array() ) {
 192                 if ( $this->isReadOnly() ) {
 193                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 194                 }
 195                 if ( empty( $opts['force'] ) ) { // sanity
 196                         unset( $opts['nonLocking'] );
 197                         unset( $opts['allowStale'] );
 198                 }
 199                 return $this->doOperationsInternal( $ops, $opts );
 200         }
 201
 202         /**
 203          * @see FileBackend::doOperations()
 204          */
 205         abstract protected function doOperationsInternal( array $ops, array $opts );
 206
 207         /**
 208          * Same as doOperations() except it takes a single operation.
 209          * If you are doing a batch of operations that should either
 210          * all succeed or all fail, then use that function instead.
 211          *
 212          * @see FileBackend::doOperations()
 213          *
 214          * @param $op Array Operation
 215          * @param $opts Array Operation options
 216          * @return Status
 217          */
 218         final public function doOperation( array $op, array $opts = array() ) {
 219                 return $this->doOperations( array( $op ), $opts );
 220         }
 221
 222         /**
 223          * Performs a single create operation.
 224          * This sets $params['op'] to 'create' and passes it to doOperation().
 225          *
 226          * @see FileBackend::doOperation()
 227          *
 228          * @param $params Array Operation parameters
 229          * @param $opts Array Operation options
 230          * @return Status
 231          */
 232         final public function create( array $params, array $opts = array() ) {
 233                 $params['op'] = 'create';
 234                 return $this->doOperation( $params, $opts );
 235         }
 236
 237         /**
 238          * Performs a single store operation.
 239          * This sets $params['op'] to 'store' and passes it to doOperation().
 240          *
 241          * @see FileBackend::doOperation()
 242          *
 243          * @param $params Array Operation parameters
 244          * @param $opts Array Operation options
 245          * @return Status
 246          */
 247         final public function store( array $params, array $opts = array() ) {
 248                 $params['op'] = 'store';
 249                 return $this->doOperation( $params, $opts );
 250         }
 251
 252         /**
 253          * Performs a single copy operation.
 254          * This sets $params['op'] to 'copy' and passes it to doOperation().
 255          *
 256          * @see FileBackend::doOperation()
 257          *
 258          * @param $params Array Operation parameters
 259          * @param $opts Array Operation options
 260          * @return Status
 261          */
 262         final public function copy( array $params, array $opts = array() ) {
 263                 $params['op'] = 'copy';
 264                 return $this->doOperation( $params, $opts );
 265         }
 266
 267         /**
 268          * Performs a single move operation.
 269          * This sets $params['op'] to 'move' and passes it to doOperation().
 270          *
 271          * @see FileBackend::doOperation()
 272          *
 273          * @param $params Array Operation parameters
 274          * @param $opts Array Operation options
 275          * @return Status
 276          */
 277         final public function move( array $params, array $opts = array() ) {
 278                 $params['op'] = 'move';
 279                 return $this->doOperation( $params, $opts );
 280         }
 281
 282         /**
 283          * Performs a single delete operation.
 284          * This sets $params['op'] to 'delete' and passes it to doOperation().
 285          *
 286          * @see FileBackend::doOperation()
 287          *
 288          * @param $params Array Operation parameters
 289          * @param $opts Array Operation options
 290          * @return Status
 291          */
 292         final public function delete( array $params, array $opts = array() ) {
 293                 $params['op'] = 'delete';
 294                 return $this->doOperation( $params, $opts );
 295         }
 296
 297         /**
 298          * Concatenate a list of storage files into a single file system file.
 299          * The target path should refer to a file that is already locked or
 300          * otherwise safe from modification from other processes. Normally,
 301          * the file will be a new temp file, which should be adequate.
 302          * $params include:
 303          *     srcs          : ordered source storage paths (e.g. chunk1, chunk2, ...)
 304          *     dst           : file system path to 0-byte temp file
 305          *
 306          * @param $params Array Operation parameters
 307          * @return Status
 308          */
 309         abstract public function concatenate( array $params );
 310
 311         /**
 312          * Prepare a storage directory for usage.
 313          * This will create any required containers and parent directories.
 314          * Backends using key/value stores only need to create the container.
 315          *
 316          * $params include:
 317          *     dir : storage directory
 318          *
 319          * @param $params Array
 320          * @return Status
 321          */
 322         final public function prepare( array $params ) {
 323                 if ( $this->isReadOnly() ) {
 324                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 325                 }
 326                 return $this->doPrepare( $params );
 327         }
 328
 329         /**
 330          * @see FileBackend::prepare()
 331          */
 332         abstract protected function doPrepare( array $params );
 333
 334         /**
 335          * Take measures to block web access to a storage directory and
 336          * the container it belongs to. FS backends might add .htaccess
 337          * files whereas key/value store backends might restrict container
 338          * access to the auth user that represents end-users in web request.
 339          * This is not guaranteed to actually do anything.
 340          *
 341          * $params include:
 342          *     dir       : storage directory
 343          *     noAccess  : try to deny file access
 344          *     noListing : try to deny file listing
 345          *
 346          * @param $params Array
 347          * @return Status
 348          */
 349         final public function secure( array $params ) {
 350                 if ( $this->isReadOnly() ) {
 351                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 352                 }
 353                 $status = $this->doPrepare( $params ); // dir must exist to restrict it
 354                 if ( $status->isOK() ) {
 355                         $status->merge( $this->doSecure( $params ) );
 356                 }
 357                 return $status;
 358         }
 359
 360         /**
 361          * @see FileBackend::secure()
 362          */
 363         abstract protected function doSecure( array $params );
 364
 365         /**
 366          * Delete a storage directory if it is empty.
 367          * Backends using key/value stores may do nothing unless the directory
 368          * is that of an empty container, in which case it should be deleted.
 369          *
 370          * $params include:
 371          *     dir : storage directory
 372          *
 373          * @param $params Array
 374          * @return Status
 375          */
 376         final public function clean( array $params ) {
 377                 if ( $this->isReadOnly() ) {
 378                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 379                 }
 380                 return $this->doClean( $params );
 381         }
 382
 383         /**
 384          * @see FileBackend::clean()
 385          */
 386         abstract protected function doClean( array $params );
 387
 388         /**
 389          * Check if a file exists at a storage path in the backend.
 390          * This returns false if only a directory exists at the path.
 391          *
 392          * $params include:
 393          *     src    : source storage path
 394          *     latest : use the latest available data
 395          *
 396          * @param $params Array
 397          * @return bool|null Returns null on failure
 398          */
 399         abstract public function fileExists( array $params );
 400
 401         /**
 402          * Get the last-modified timestamp of the file at a storage path.
 403          *
 404          * $params include:
 405          *     src    : source storage path
 406          *     latest : use the latest available data
 407          *
 408          * @param $params Array
 409          * @return string|false TS_MW timestamp or false on failure
 410          */
 411         abstract public function getFileTimestamp( array $params );
 412
 413         /**
 414          * Get the contents of a file at a storage path in the backend.
 415          * This should be avoided for potentially large files.
 416          *
 417          * $params include:
 418          *     src    : source storage path
 419          *     latest : use the latest available data
 420          *
 421          * @param $params Array
 422          * @return string|false Returns false on failure
 423          */
 424         abstract public function getFileContents( array $params );
 425
 426         /**
 427          * Get the size (bytes) of a file at a storage path in the backend.
 428          *
 429          * $params include:
 430          *     src    : source storage path
 431          *     latest : use the latest available data
 432          *
 433          * @param $params Array
 434          * @return integer|false Returns false on failure
 435          */
 436         abstract public function getFileSize( array $params );
 437
 438         /**
 439          * Get quick information about a file at a storage path in the backend.
 440          * If the file does not exist, then this returns false.
 441          * Otherwise, the result is an associative array that includes:
 442          *     mtime  : the last-modified timestamp (TS_MW)
 443          *     size   : the file size (bytes)
 444          * Additional values may be included for internal use only.
 445          *
 446          * $params include:
 447          *     src    : source storage path
 448          *     latest : use the latest available data
 449          *
 450          * @param $params Array
 451          * @return Array|false|null Returns null on failure
 452          */
 453         abstract public function getFileStat( array $params );
 454
 455         /**
 456          * Get a SHA-1 hash of the file at a storage path in the backend.
 457          *
 458          * $params include:
 459          *     src    : source storage path
 460          *     latest : use the latest available data
 461          *
 462          * @param $params Array
 463          * @return string|false Hash string or false on failure
 464          */
 465         abstract public function getFileSha1Base36( array $params );
 466
 467         /**
 468          * Get the properties of the file at a storage path in the backend.
 469          * Returns FSFile::placeholderProps() on failure.
 470          *
 471          * $params include:
 472          *     src    : source storage path
 473          *     latest : use the latest available data
 474          *
 475          * @param $params Array
 476          * @return Array
 477          */
 478         abstract public function getFileProps( array $params );
 479
 480         /**
 481          * Stream the file at a storage path in the backend.
 482          * If the file does not exists, a 404 error will be given.
 483          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 484          * must be sent if streaming began, while none should be sent otherwise.
 485          * Implementations should flush the output buffer before sending data.
 486          *
 487          * $params include:
 488          *     src     : source storage path
 489          *     headers : additional HTTP headers to send on success
 490          *     latest  : use the latest available data
 491          *
 492          * @param $params Array
 493          * @return Status
 494          */
 495         abstract public function streamFile( array $params );
 496
 497         /**
 498          * Returns a file system file, identical to the file at a storage path.
 499          * The file returned is either:
 500          * a) A local copy of the file at a storage path in the backend.
 501          *    The temporary copy will have the same extension as the source.
 502          * b) An original of the file at a storage path in the backend.
 503          * Temporary files may be purged when the file object falls out of scope.
 504          *
 505          * Write operations should *never* be done on this file as some backends
 506          * may do internal tracking or may be instances of FileBackendMultiWrite.
 507          * In that later case, there are copies of the file that must stay in sync.
 508          * Additionally, further calls to this function may return the same file.
 509          *
 510          * $params include:
 511          *     src    : source storage path
 512          *     latest : use the latest available data
 513          *
 514          * @param $params Array
 515          * @return FSFile|null Returns null on failure
 516          */
 517         abstract public function getLocalReference( array $params );
 518
 519         /**
 520          * Get a local copy on disk of the file at a storage path in the backend.
 521          * The temporary copy will have the same file extension as the source.
 522          * Temporary files may be purged when the file object falls out of scope.
 523          *
 524          * $params include:
 525          *     src    : source storage path
 526          *     latest : use the latest available data
 527          *
 528          * @param $params Array
 529          * @return TempFSFile|null Returns null on failure
 530          */
 531         abstract public function getLocalCopy( array $params );
 532
 533         /**
 534          * Get an iterator to list out all stored files under a storage directory.
 535          * If the directory is of the form "mwstore://backend/container",
 536          * then all files in the container should be listed.
 537          * If the directory is of form "mwstore://backend/container/dir",
 538          * then all files under that container directory should be listed.
 539          * Results should be storage paths relative to the given directory.
 540          *
 541          * Storage backends with eventual consistency might return stale data.
 542          *
 543          * $params include:
 544          *     dir : storage path directory
 545          *
 546          * @return Traversable|Array|null Returns null on failure
 547          */
 548         abstract public function getFileList( array $params );
 549
 550         /**
 551          * Invalidate any in-process file existence and property cache.
 552          * If $paths is given, then only the cache for those files will be cleared.
 553          *
 554          * @param $paths Array Storage paths (optional)
 555          * @return void
 556          */
 557         public function clearCache( array $paths = null ) {}
 558
 559         /**
 560          * Lock the files at the given storage paths in the backend.
 561          * This will either lock all the files or none (on failure).
 562          *
 563          * Callers should consider using getScopedFileLocks() instead.
 564          *
 565          * @param $paths Array Storage paths
 566          * @param $type integer LockManager::LOCK_* constant
 567          * @return Status
 568          */
 569         final public function lockFiles( array $paths, $type ) {
 570                 return $this->lockManager->lock( $paths, $type );
 571         }
 572
 573         /**
 574          * Unlock the files at the given storage paths in the backend.
 575          *
 576          * @param $paths Array Storage paths
 577          * @param $type integer LockManager::LOCK_* constant
 578          * @return Status
 579          */
 580         final public function unlockFiles( array $paths, $type ) {
 581                 return $this->lockManager->unlock( $paths, $type );
 582         }
 583
 584         /**
 585          * Lock the files at the given storage paths in the backend.
 586          * This will either lock all the files or none (on failure).
 587          * On failure, the status object will be updated with errors.
 588          *
 589          * Once the return value goes out scope, the locks will be released and
 590          * the status updated. Unlock fatals will not change the status "OK" value.
 591          *
 592          * @param $paths Array Storage paths
 593          * @param $type integer LockManager::LOCK_* constant
 594          * @param $status Status Status to update on lock/unlock
 595          * @return ScopedLock|null Returns null on failure
 596          */
 597         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
 598                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
 599         }
 600
 601         /**
 602          * Check if a given path is a "mwstore://" path.
 603          * This does not do any further validation or any existence checks.
 604          *
 605          * @param $path string
 606          * @return bool
 607          */
 608         final public static function isStoragePath( $path ) {
 609                 return ( strpos( $path, 'mwstore://' ) === 0 );
 610         }
 611
 612         /**
 613          * Split a storage path into a backend name, a container name,
 614          * and a relative file path. The relative path may be the empty string.
 615          * This does not do any path normalization or traversal checks.
 616          *
 617          * @param $storagePath string
 618          * @return Array (backend, container, rel object) or (null, null, null)
 619          */
 620         final public static function splitStoragePath( $storagePath ) {
 621                 if ( self::isStoragePath( $storagePath ) ) {
 622                         // Remove the "mwstore://" prefix and split the path
 623                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
 624                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
 625                                 if ( count( $parts ) == 3 ) {
 626                                         return $parts; // e.g. "backend/container/path"
 627                                 } else {
 628                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
 629                                 }
 630                         }
 631                 }
 632                 return array( null, null, null );
 633         }
 634
 635         /**
 636          * Normalize a storage path by cleaning up directory separators.
 637          * Returns null if the path is not of the format of a valid storage path.
 638          *
 639          * @param $storagePath string
 640          * @return string|null
 641          */
 642         final public static function normalizeStoragePath( $storagePath ) {
 643                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
 644                 if ( $relPath !== null ) { // must be for this backend
 645                         $relPath = self::normalizeContainerPath( $relPath );
 646                         if ( $relPath !== null ) {
 647                                 return ( $relPath != '' )
 648                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
 649                                         : "mwstore://{$backend}/{$container}";
 650                         }
 651                 }
 652                 return null;
 653         }
 654
 655         /**
 656          * Validate and normalize a relative storage path.
 657          * Null is returned if the path involves directory traversal.
 658          * Traversal is insecure for FS backends and broken for others.
 659          *
 660          * @param $path string Storage path relative to a container
 661          * @return string|null
 662          */
 663         final protected static function normalizeContainerPath( $path ) {
 664                 // Normalize directory separators
 665                 $path = strtr( $path, '\\', '/' );
 666                 // Collapse any consecutive directory separators
 667                 $path = preg_replace( '![/]{2,}!', '/', $path );
 668                 // Remove any leading directory separator
 669                 $path = ltrim( $path, '/' );
 670                 // Use the same traversal protection as Title::secureAndSplit()
 671                 if ( strpos( $path, '.' ) !== false ) {
 672                         if (
 673                                 $path === '.' ||
 674                                 $path === '..' ||
 675                                 strpos( $path, './' ) === 0 ||
 676                                 strpos( $path, '../' ) === 0 ||
 677                                 strpos( $path, '/./' ) !== false ||
 678                                 strpos( $path, '/../' ) !== false
 679                         ) {
 680                                 return null;
 681                         }
 682                 }
 683                 return $path;
 684         }
 685
 686         /**
 687          * Get the parent storage directory of a storage path.
 688          * This returns a path like "mwstore://backend/container",
 689          * "mwstore://backend/container/...", or null if there is no parent.
 690          *
 691          * @param $storagePath string
 692          * @return string|null
 693          */
 694         final public static function parentStoragePath( $storagePath ) {
 695                 $storagePath = dirname( $storagePath );
 696                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
 697                 return ( $rel === null ) ? null : $storagePath;
 698         }
 699
 700         /**
 701          * Get the final extension from a storage or FS path
 702          *
 703          * @param $path string
 704          * @return string
 705          */
 706         final public static function extensionFromPath( $path ) {
 707                 $i = strrpos( $path, '.' );
 708                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
 709         }
 710 }
 711
 712 /**
 713  * @brief Base class for all backends associated with a particular storage medium.
 714  *
 715  * This class defines the methods as abstract that subclasses must implement.
 716  * Outside callers should *not* use functions with "Internal" in the name.
 717  *
 718  * The FileBackend operations are implemented using basic functions
 719  * such as storeInternal(), copyInternal(), deleteInternal() and the like.
 720  * This class is also responsible for path resolution and sanitization.
 721  *
 722  * @ingroup FileBackend
 723  * @since 1.19
 724  */
 725 abstract class FileBackendStore extends FileBackend {
 726         /** @var Array Map of paths to small (RAM/disk) cache items */
 727         protected $cache = array(); // (storage path => key => value)
 728         protected $maxCacheSize = 100; // integer; max paths with entries
 729         /** @var Array Map of paths to large (RAM/disk) cache items */
 730         protected $expensiveCache = array(); // (storage path => key => value)
 731         protected $maxExpensiveCacheSize = 10; // integer; max paths with entries
 732
 733         /** @var Array Map of container names to sharding settings */
 734         protected $shardViaHashLevels = array(); // (container name => config array)
 735
 736         protected $maxFileSize = 1000000000; // integer bytes (1GB)
 737
 738         /**
 739          * Get the maximum allowable file size given backend
 740          * medium restrictions and basic performance constraints.
 741          * Do not call this function from places outside FileBackend and FileOp.
 742          *
 743          * @return integer Bytes
 744          */
 745         final public function maxFileSizeInternal() {
 746                 return $this->maxFileSize;
 747         }
 748
 749         /**
 750          * Check if a file can be created at a given storage path.
 751          * FS backends should check if the parent directory exists and the file is writable.
 752          * Backends using key/value stores should check if the container exists.
 753          *
 754          * @param $storagePath string
 755          * @return bool
 756          */
 757         abstract public function isPathUsableInternal( $storagePath );
 758
 759         /**
 760          * Create a file in the backend with the given contents.
 761          * Do not call this function from places outside FileBackend and FileOp.
 762          *
 763          * $params include:
 764          *     content       : the raw file contents
 765          *     dst           : destination storage path
 766          *     overwrite     : overwrite any file that exists at the destination
 767          *
 768          * @param $params Array
 769          * @return Status
 770          */
 771         final public function createInternal( array $params ) {
 772                 wfProfileIn( __METHOD__ );
 773                 if ( strlen( $params['content'] ) > $this->maxFileSizeInternal() ) {
 774                         $status = Status::newFatal( 'backend-fail-create', $params['dst'] );
 775                 } else {
 776                         $status = $this->doCreateInternal( $params );
 777                         $this->clearCache( array( $params['dst'] ) );
 778                 }
 779                 wfProfileOut( __METHOD__ );
 780                 return $status;
 781         }
 782
 783         /**
 784          * @see FileBackendStore::createInternal()
 785          */
 786         abstract protected function doCreateInternal( array $params );
 787
 788         /**
 789          * Store a file into the backend from a file on disk.
 790          * Do not call this function from places outside FileBackend and FileOp.
 791          *
 792          * $params include:
 793          *     src           : source path on disk
 794          *     dst           : destination storage path
 795          *     overwrite     : overwrite any file that exists at the destination
 796          *
 797          * @param $params Array
 798          * @return Status
 799          */
 800         final public function storeInternal( array $params ) {
 801                 wfProfileIn( __METHOD__ );
 802                 if ( filesize( $params['src'] ) > $this->maxFileSizeInternal() ) {
 803                         $status = Status::newFatal( 'backend-fail-store', $params['dst'] );
 804                 } else {
 805                         $status = $this->doStoreInternal( $params );
 806                         $this->clearCache( array( $params['dst'] ) );
 807                 }
 808                 wfProfileOut( __METHOD__ );
 809                 return $status;
 810         }
 811
 812         /**
 813          * @see FileBackendStore::storeInternal()
 814          */
 815         abstract protected function doStoreInternal( array $params );
 816
 817         /**
 818          * Copy a file from one storage path to another in the backend.
 819          * Do not call this function from places outside FileBackend and FileOp.
 820          *
 821          * $params include:
 822          *     src           : source storage path
 823          *     dst           : destination storage path
 824          *     overwrite     : overwrite any file that exists at the destination
 825          *
 826          * @param $params Array
 827          * @return Status
 828          */
 829         final public function copyInternal( array $params ) {
 830                 wfProfileIn( __METHOD__ );
 831                 $status = $this->doCopyInternal( $params );
 832                 $this->clearCache( array( $params['dst'] ) );
 833                 wfProfileOut( __METHOD__ );
 834                 return $status;
 835         }
 836
 837         /**
 838          * @see FileBackendStore::copyInternal()
 839          */
 840         abstract protected function doCopyInternal( array $params );
 841
 842         /**
 843          * Delete a file at the storage path.
 844          * Do not call this function from places outside FileBackend and FileOp.
 845          *
 846          * $params include:
 847          *     src                 : source storage path
 848          *     ignoreMissingSource : do nothing if the source file does not exist
 849          *
 850          * @param $params Array
 851          * @return Status
 852          */
 853         final public function deleteInternal( array $params ) {
 854                 wfProfileIn( __METHOD__ );
 855                 $status = $this->doDeleteInternal( $params );
 856                 $this->clearCache( array( $params['src'] ) );
 857                 wfProfileOut( __METHOD__ );
 858                 return $status;
 859         }
 860
 861         /**
 862          * @see FileBackendStore::deleteInternal()
 863          */
 864         abstract protected function doDeleteInternal( array $params );
 865
 866         /**
 867          * Move a file from one storage path to another in the backend.
 868          * Do not call this function from places outside FileBackend and FileOp.
 869          *
 870          * $params include:
 871          *     src           : source storage path
 872          *     dst           : destination storage path
 873          *     overwrite     : overwrite any file that exists at the destination
 874          *
 875          * @param $params Array
 876          * @return Status
 877          */
 878         final public function moveInternal( array $params ) {
 879                 wfProfileIn( __METHOD__ );
 880                 $status = $this->doMoveInternal( $params );
 881                 $this->clearCache( array( $params['src'], $params['dst'] ) );
 882                 wfProfileOut( __METHOD__ );
 883                 return $status;
 884         }
 885
 886         /**
 887          * @see FileBackendStore::moveInternal()
 888          */
 889         protected function doMoveInternal( array $params ) {
 890                 // Copy source to dest
 891                 $status = $this->copyInternal( $params );
 892                 if ( $status->isOK() ) {
 893                         // Delete source (only fails due to races or medium going down)
 894                         $status->merge( $this->deleteInternal( array( 'src' => $params['src'] ) ) );
 895                         $status->setResult( true, $status->value ); // ignore delete() errors
 896                 }
 897                 return $status;
 898         }
 899
 900         /**
 901          * @see FileBackend::concatenate()
 902          */
 903         final public function concatenate( array $params ) {
 904                 wfProfileIn( __METHOD__ );
 905                 $status = Status::newGood();
 906
 907                 // Try to lock the source files for the scope of this function
 908                 $scopeLockS = $this->getScopedFileLocks( $params['srcs'], LockManager::LOCK_UW, $status );
 909                 if ( $status->isOK() ) {
 910                         // Actually do the concatenation
 911                         $status->merge( $this->doConcatenate( $params ) );
 912                 }
 913
 914                 wfProfileOut( __METHOD__ );
 915                 return $status;
 916         }
 917
 918         /**
 919          * @see FileBackendStore::concatenate()
 920          */
 921         protected function doConcatenate( array $params ) {
 922                 $status = Status::newGood();
 923                 $tmpPath = $params['dst']; // convenience
 924
 925                 // Check that the specified temp file is valid...
 926                 wfSuppressWarnings();
 927                 $ok = ( is_file( $tmpPath ) && !filesize( $tmpPath ) );
 928                 wfRestoreWarnings();
 929                 if ( !$ok ) { // not present or not empty
 930                         $status->fatal( 'backend-fail-opentemp', $tmpPath );
 931                         return $status;
 932                 }
 933
 934                 // Build up the temp file using the source chunks (in order)...
 935                 $tmpHandle = fopen( $tmpPath, 'ab' );
 936                 if ( $tmpHandle === false ) {
 937                         $status->fatal( 'backend-fail-opentemp', $tmpPath );
 938                         return $status;
 939                 }
 940                 foreach ( $params['srcs'] as $virtualSource ) {
 941                         // Get a local FS version of the chunk
 942                         $tmpFile = $this->getLocalReference( array( 'src' => $virtualSource ) );
 943                         if ( !$tmpFile ) {
 944                                 $status->fatal( 'backend-fail-read', $virtualSource );
 945                                 return $status;
 946                         }
 947                         // Get a handle to the local FS version
 948                         $sourceHandle = fopen( $tmpFile->getPath(), 'r' );
 949                         if ( $sourceHandle === false ) {
 950                                 fclose( $tmpHandle );
 951                                 $status->fatal( 'backend-fail-read', $virtualSource );
 952                                 return $status;
 953                         }
 954                         // Append chunk to file (pass chunk size to avoid magic quotes)
 955                         if ( !stream_copy_to_stream( $sourceHandle, $tmpHandle ) ) {
 956                                 fclose( $sourceHandle );
 957                                 fclose( $tmpHandle );
 958                                 $status->fatal( 'backend-fail-writetemp', $tmpPath );
 959                                 return $status;
 960                         }
 961                         fclose( $sourceHandle );
 962                 }
 963                 if ( !fclose( $tmpHandle ) ) {
 964                         $status->fatal( 'backend-fail-closetemp', $tmpPath );
 965                         return $status;
 966                 }
 967
 968                 clearstatcache(); // temp file changed
 969
 970                 return $status;
 971         }
 972
 973         /**
 974          * @see FileBackend::doPrepare()
 975          */
 976         final protected function doPrepare( array $params ) {
 977                 wfProfileIn( __METHOD__ );
 978
 979                 $status = Status::newGood();
 980                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
 981                 if ( $dir === null ) {
 982                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
 983                         wfProfileOut( __METHOD__ );
 984                         return $status; // invalid storage path
 985                 }
 986
 987                 if ( $shard !== null ) { // confined to a single container/shard
 988                         $status->merge( $this->doPrepareInternal( $fullCont, $dir, $params ) );
 989                 } else { // directory is on several shards
 990                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
 991                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
 992                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
 993                                 $status->merge( $this->doPrepareInternal( "{$fullCont}{$suffix}", $dir, $params ) );
 994                         }
 995                 }
 996
 997                 wfProfileOut( __METHOD__ );
 998                 return $status;
 999         }
1000
1001         /**
1002          * @see FileBackendStore::doPrepare()
1003          */
1004         protected function doPrepareInternal( $container, $dir, array $params ) {
1005                 return Status::newGood();
1006         }
1007
1008         /**
1009          * @see FileBackend::doSecure()
1010          */
1011         final protected function doSecure( array $params ) {
1012                 wfProfileIn( __METHOD__ );
1013                 $status = Status::newGood();
1014
1015                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
1016                 if ( $dir === null ) {
1017                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
1018                         wfProfileOut( __METHOD__ );
1019                         return $status; // invalid storage path
1020                 }
1021
1022                 if ( $shard !== null ) { // confined to a single container/shard
1023                         $status->merge( $this->doSecureInternal( $fullCont, $dir, $params ) );
1024                 } else { // directory is on several shards
1025                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
1026                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
1027                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
1028                                 $status->merge( $this->doSecureInternal( "{$fullCont}{$suffix}", $dir, $params ) );
1029                         }
1030                 }
1031
1032                 wfProfileOut( __METHOD__ );
1033                 return $status;
1034         }
1035
1036         /**
1037          * @see FileBackendStore::doSecure()
1038          */
1039         protected function doSecureInternal( $container, $dir, array $params ) {
1040                 return Status::newGood();
1041         }
1042
1043         /**
1044          * @see FileBackend::doClean()
1045          */
1046         final protected function doClean( array $params ) {
1047                 wfProfileIn( __METHOD__ );
1048                 $status = Status::newGood();
1049
1050                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
1051                 if ( $dir === null ) {
1052                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
1053                         wfProfileOut( __METHOD__ );
1054                         return $status; // invalid storage path
1055                 }
1056
1057                 // Attempt to lock this directory...
1058                 $filesLockEx = array( $params['dir'] );
1059                 $scopedLockE = $this->getScopedFileLocks( $filesLockEx, LockManager::LOCK_EX, $status );
1060                 if ( !$status->isOK() ) {
1061                         wfProfileOut( __METHOD__ );
1062                         return $status; // abort
1063                 }
1064
1065                 if ( $shard !== null ) { // confined to a single container/shard
1066                         $status->merge( $this->doCleanInternal( $fullCont, $dir, $params ) );
1067                 } else { // directory is on several shards
1068                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
1069                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
1070                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
1071                                 $status->merge( $this->doCleanInternal( "{$fullCont}{$suffix}", $dir, $params ) );
1072                         }
1073                 }
1074
1075                 wfProfileOut( __METHOD__ );
1076                 return $status;
1077         }
1078
1079         /**
1080          * @see FileBackendStore::doClean()
1081          */
1082         protected function doCleanInternal( $container, $dir, array $params ) {
1083                 return Status::newGood();
1084         }
1085
1086         /**
1087          * @see FileBackend::fileExists()
1088          */
1089         final public function fileExists( array $params ) {
1090                 wfProfileIn( __METHOD__ );
1091                 $stat = $this->getFileStat( $params );
1092                 wfProfileOut( __METHOD__ );
1093                 return ( $stat === null ) ? null : (bool)$stat; // null => failure
1094         }
1095
1096         /**
1097          * @see FileBackend::getFileTimestamp()
1098          */
1099         final public function getFileTimestamp( array $params ) {
1100                 wfProfileIn( __METHOD__ );
1101                 $stat = $this->getFileStat( $params );
1102                 wfProfileOut( __METHOD__ );
1103                 return $stat ? $stat['mtime'] : false;
1104         }
1105
1106         /**
1107          * @see FileBackend::getFileSize()
1108          */
1109         final public function getFileSize( array $params ) {
1110                 wfProfileIn( __METHOD__ );
1111                 $stat = $this->getFileStat( $params );
1112                 wfProfileOut( __METHOD__ );
1113                 return $stat ? $stat['size'] : false;
1114         }
1115
1116         /**
1117          * @see FileBackend::getFileStat()
1118          */
1119         final public function getFileStat( array $params ) {
1120                 wfProfileIn( __METHOD__ );
1121                 $path = self::normalizeStoragePath( $params['src'] );
1122                 if ( $path === null ) {
1123                         return false; // invalid storage path
1124                 }
1125                 $latest = !empty( $params['latest'] );
1126                 if ( isset( $this->cache[$path]['stat'] ) ) {
1127                         // If we want the latest data, check that this cached
1128                         // value was in fact fetched with the latest available data.
1129                         if ( !$latest || $this->cache[$path]['stat']['latest'] ) {
1130                                 wfProfileOut( __METHOD__ );
1131                                 return $this->cache[$path]['stat'];
1132                         }
1133                 }
1134                 $stat = $this->doGetFileStat( $params );
1135                 if ( is_array( $stat ) ) { // don't cache negatives
1136                         $this->trimCache(); // limit memory
1137                         $this->cache[$path]['stat'] = $stat;
1138                         $this->cache[$path]['stat']['latest'] = $latest;
1139                 }
1140                 wfProfileOut( __METHOD__ );
1141                 return $stat;
1142         }
1143
1144         /**
1145          * @see FileBackendStore::getFileStat()
1146          */
1147         abstract protected function doGetFileStat( array $params );
1148
1149         /**
1150          * @see FileBackend::getFileContents()
1151          */
1152         public function getFileContents( array $params ) {
1153                 wfProfileIn( __METHOD__ );
1154                 $tmpFile = $this->getLocalReference( $params );
1155                 if ( !$tmpFile ) {
1156                         wfProfileOut( __METHOD__ );
1157                         return false;
1158                 }
1159                 wfSuppressWarnings();
1160                 $data = file_get_contents( $tmpFile->getPath() );
1161                 wfRestoreWarnings();
1162                 wfProfileOut( __METHOD__ );
1163                 return $data;
1164         }
1165
1166         /**
1167          * @see FileBackend::getFileSha1Base36()
1168          */
1169         final public function getFileSha1Base36( array $params ) {
1170                 wfProfileIn( __METHOD__ );
1171                 $path = $params['src'];
1172                 if ( isset( $this->cache[$path]['sha1'] ) ) {
1173                         wfProfileOut( __METHOD__ );
1174                         return $this->cache[$path]['sha1'];
1175                 }
1176                 $hash = $this->doGetFileSha1Base36( $params );
1177                 if ( $hash ) { // don't cache negatives
1178                         $this->trimCache(); // limit memory
1179                         $this->cache[$path]['sha1'] = $hash;
1180                 }
1181                 wfProfileOut( __METHOD__ );
1182                 return $hash;
1183         }
1184
1185         /**
1186          * @see FileBackendStore::getFileSha1Base36()
1187          */
1188         protected function doGetFileSha1Base36( array $params ) {
1189                 $fsFile = $this->getLocalReference( $params );
1190                 if ( !$fsFile ) {
1191                         return false;
1192                 } else {
1193                         return $fsFile->getSha1Base36();
1194                 }
1195         }
1196
1197         /**
1198          * @see FileBackend::getFileProps()
1199          */
1200         final public function getFileProps( array $params ) {
1201                 wfProfileIn( __METHOD__ );
1202                 $fsFile = $this->getLocalReference( $params );
1203                 $props = $fsFile ? $fsFile->getProps() : FSFile::placeholderProps();
1204                 wfProfileOut( __METHOD__ );
1205                 return $props;
1206         }
1207
1208         /**
1209          * @see FileBackend::getLocalReference()
1210          */
1211         public function getLocalReference( array $params ) {
1212                 wfProfileIn( __METHOD__ );
1213                 $path = $params['src'];
1214                 if ( isset( $this->expensiveCache[$path]['localRef'] ) ) {
1215                         wfProfileOut( __METHOD__ );
1216                         return $this->expensiveCache[$path]['localRef'];
1217                 }
1218                 $tmpFile = $this->getLocalCopy( $params );
1219                 if ( $tmpFile ) { // don't cache negatives
1220                         $this->trimExpensiveCache(); // limit memory
1221                         $this->expensiveCache[$path]['localRef'] = $tmpFile;
1222                 }
1223                 wfProfileOut( __METHOD__ );
1224                 return $tmpFile;
1225         }
1226
1227         /**
1228          * @see FileBackend::streamFile()
1229          */
1230         final public function streamFile( array $params ) {
1231                 wfProfileIn( __METHOD__ );
1232                 $status = Status::newGood();
1233
1234                 $info = $this->getFileStat( $params );
1235                 if ( !$info ) { // let StreamFile handle the 404
1236                         $status->fatal( 'backend-fail-notexists', $params['src'] );
1237                 }
1238
1239                 // Set output buffer and HTTP headers for stream
1240                 $extraHeaders = $params['headers'] ? $params['headers'] : array();
1241                 $res = StreamFile::prepareForStream( $params['src'], $info, $extraHeaders );
1242                 if ( $res == StreamFile::NOT_MODIFIED ) {
1243                         // do nothing; client cache is up to date
1244                 } elseif ( $res == StreamFile::READY_STREAM ) {
1245                         $status = $this->doStreamFile( $params );
1246                 } else {
1247                         $status->fatal( 'backend-fail-stream', $params['src'] );
1248                 }
1249
1250                 wfProfileOut( __METHOD__ );
1251                 return $status;
1252         }
1253
1254         /**
1255          * @see FileBackendStore::streamFile()
1256          */
1257         protected function doStreamFile( array $params ) {
1258                 $status = Status::newGood();
1259
1260                 $fsFile = $this->getLocalReference( $params );
1261                 if ( !$fsFile ) {
1262                         $status->fatal( 'backend-fail-stream', $params['src'] );
1263                 } elseif ( !readfile( $fsFile->getPath() ) ) {
1264                         $status->fatal( 'backend-fail-stream', $params['src'] );
1265                 }
1266
1267                 return $status;
1268         }
1269
1270         /**
1271          * @copydoc FileBackend::getFileList()
1272          */
1273         final public function getFileList( array $params ) {
1274                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
1275                 if ( $dir === null ) { // invalid storage path
1276                         return null;
1277                 }
1278                 if ( $shard !== null ) {
1279                         // File listing is confined to a single container/shard
1280                         return $this->getFileListInternal( $fullCont, $dir, $params );
1281                 } else {
1282                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
1283                         // File listing spans multiple containers/shards
1284                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
1285                         return new FileBackendStoreShardListIterator( $this,
1286                                 $fullCont, $dir, $this->getContainerSuffixes( $shortCont ), $params );
1287                 }
1288         }
1289
1290         /**
1291          * Do not call this function from places outside FileBackend
1292          *
1293          * @see FileBackendStore::getFileList()
1294          *
1295          * @param $container string Resolved container name
1296          * @param $dir string Resolved path relative to container
1297          * @param $params Array
1298          * @return Traversable|Array|null
1299          */
1300         abstract public function getFileListInternal( $container, $dir, array $params );
1301
1302         /**
1303          * Get the list of supported operations and their corresponding FileOp classes.
1304          *
1305          * @return Array
1306          */
1307         protected function supportedOperations() {
1308                 return array(
1309                         'store'       => 'StoreFileOp',
1310                         'copy'        => 'CopyFileOp',
1311                         'move'        => 'MoveFileOp',
1312                         'delete'      => 'DeleteFileOp',
1313                         'create'      => 'CreateFileOp',
1314                         'null'        => 'NullFileOp'
1315                 );
1316         }
1317
1318         /**
1319          * Return a list of FileOp objects from a list of operations.
1320          * Do not call this function from places outside FileBackend.
1321          *
1322          * The result must have the same number of items as the input.
1323          * An exception is thrown if an unsupported operation is requested.
1324          *
1325          * @param $ops Array Same format as doOperations()
1326          * @return Array List of FileOp objects
1327          * @throws MWException
1328          */
1329         final public function getOperations( array $ops ) {
1330                 $supportedOps = $this->supportedOperations();
1331
1332                 $performOps = array(); // array of FileOp objects
1333                 // Build up ordered array of FileOps...
1334                 foreach ( $ops as $operation ) {
1335                         $opName = $operation['op'];
1336                         if ( isset( $supportedOps[$opName] ) ) {
1337                                 $class = $supportedOps[$opName];
1338                                 // Get params for this operation
1339                                 $params = $operation;
1340                                 // Append the FileOp class
1341                                 $performOps[] = new $class( $this, $params );
1342                         } else {
1343                                 throw new MWException( "Operation `$opName` is not supported." );
1344                         }
1345                 }
1346
1347                 return $performOps;
1348         }
1349
1350         /**
1351          * @see FileBackend::doOperationsInternal()
1352          */
1353         protected function doOperationsInternal( array $ops, array $opts ) {
1354                 wfProfileIn( __METHOD__ );
1355                 $status = Status::newGood();
1356
1357                 // Build up a list of FileOps...
1358                 $performOps = $this->getOperations( $ops );
1359
1360                 // Acquire any locks as needed...
1361                 if ( empty( $opts['nonLocking'] ) ) {
1362                         // Build up a list of files to lock...
1363                         $filesLockEx = $filesLockSh = array();
1364                         foreach ( $performOps as $fileOp ) {
1365                                 $filesLockSh = array_merge( $filesLockSh, $fileOp->storagePathsRead() );
1366                                 $filesLockEx = array_merge( $filesLockEx, $fileOp->storagePathsChanged() );
1367                         }
1368                         // Optimization: if doing an EX lock anyway, don't also set an SH one
1369                         $filesLockSh = array_diff( $filesLockSh, $filesLockEx );
1370                         // Get a shared lock on the parent directory of each path changed
1371                         $filesLockSh = array_merge( $filesLockSh, array_map( 'dirname', $filesLockEx ) );
1372                         // Try to lock those files for the scope of this function...
1373                         $scopeLockS = $this->getScopedFileLocks( $filesLockSh, LockManager::LOCK_UW, $status );
1374                         $scopeLockE = $this->getScopedFileLocks( $filesLockEx, LockManager::LOCK_EX, $status );
1375                         if ( !$status->isOK() ) {
1376                                 wfProfileOut( __METHOD__ );
1377                                 return $status; // abort
1378                         }
1379                 }
1380
1381                 // Clear any cache entries (after locks acquired)
1382                 $this->clearCache();
1383
1384                 // Actually attempt the operation batch...
1385                 $subStatus = FileOp::attemptBatch( $performOps, $opts );
1386
1387                 // Merge errors into status fields
1388                 $status->merge( $subStatus );
1389                 $status->success = $subStatus->success; // not done in merge()
1390
1391                 wfProfileOut( __METHOD__ );
1392                 return $status;
1393         }
1394
1395         /**
1396          * @see FileBackend::clearCache()
1397          */
1398         final public function clearCache( array $paths = null ) {
1399                 if ( is_array( $paths ) ) {
1400                         $paths = array_map( 'FileBackend::normalizeStoragePath', $paths );
1401                         $paths = array_filter( $paths, 'strlen' ); // remove nulls
1402                 }
1403                 if ( $paths === null ) {
1404                         $this->cache = array();
1405                         $this->expensiveCache = array();
1406                 } else {
1407                         foreach ( $paths as $path ) {
1408                                 unset( $this->cache[$path] );
1409                                 unset( $this->expensiveCache[$path] );
1410                         }
1411                 }
1412                 $this->doClearCache( $paths );
1413         }
1414
1415         /**
1416          * Clears any additional stat caches for storage paths
1417          *
1418          * @see FileBackend::clearCache()
1419          *
1420          * @param $paths Array Storage paths (optional)
1421          * @return void
1422          */
1423         protected function doClearCache( array $paths = null ) {}
1424
1425         /**
1426          * Prune the inexpensive cache if it is too big to add an item
1427          *
1428          * @return void
1429          */
1430         protected function trimCache() {
1431                 if ( count( $this->cache ) >= $this->maxCacheSize ) {
1432                         reset( $this->cache );
1433                         unset( $this->cache[key( $this->cache )] );
1434                 }
1435         }
1436
1437         /**
1438          * Prune the expensive cache if it is too big to add an item
1439          *
1440          * @return void
1441          */
1442         protected function trimExpensiveCache() {
1443                 if ( count( $this->expensiveCache ) >= $this->maxExpensiveCacheSize ) {
1444                         reset( $this->expensiveCache );
1445                         unset( $this->expensiveCache[key( $this->expensiveCache )] );
1446                 }
1447         }
1448
1449         /**
1450          * Check if a container name is valid.
1451          * This checks for for length and illegal characters.
1452          *
1453          * @param $container string
1454          * @return bool
1455          */
1456         final protected static function isValidContainerName( $container ) {
1457                 // This accounts for Swift and S3 restrictions while leaving room
1458                 // for things like '.xxx' (hex shard chars) or '.seg' (segments).
1459                 // This disallows directory separators or traversal characters.
1460                 // Note that matching strings URL encode to the same string;
1461                 // in Swift, the length restriction is *after* URL encoding.
1462                 return preg_match( '/^[a-z0-9][a-z0-9-_]{0,199}$/i', $container );
1463         }
1464
1465         /**
1466          * Splits a storage path into an internal container name,
1467          * an internal relative file name, and a container shard suffix.
1468          * Any shard suffix is already appended to the internal container name.
1469          * This also checks that the storage path is valid and within this backend.
1470          *
1471          * If the container is sharded but a suffix could not be determined,
1472          * this means that the path can only refer to a directory and can only
1473          * be scanned by looking in all the container shards.
1474          *
1475          * @param $storagePath string
1476          * @return Array (container, path, container suffix) or (null, null, null) if invalid
1477          */
1478         final protected function resolveStoragePath( $storagePath ) {
1479                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1480                 if ( $backend === $this->name ) { // must be for this backend
1481                         $relPath = self::normalizeContainerPath( $relPath );
1482                         if ( $relPath !== null ) {
1483                                 // Get shard for the normalized path if this container is sharded
1484                                 $cShard = $this->getContainerShard( $container, $relPath );
1485                                 // Validate and sanitize the relative path (backend-specific)
1486                                 $relPath = $this->resolveContainerPath( $container, $relPath );
1487                                 if ( $relPath !== null ) {
1488                                         // Prepend any wiki ID prefix to the container name
1489                                         $container = $this->fullContainerName( $container );
1490                                         if ( self::isValidContainerName( $container ) ) {
1491                                                 // Validate and sanitize the container name (backend-specific)
1492                                                 $container = $this->resolveContainerName( "{$container}{$cShard}" );
1493                                                 if ( $container !== null ) {
1494                                                         return array( $container, $relPath, $cShard );
1495                                                 }
1496                                         }
1497                                 }
1498                         }
1499                 }
1500                 return array( null, null, null );
1501         }
1502
1503         /**
1504          * Like resolveStoragePath() except null values are returned if
1505          * the container is sharded and the shard could not be determined.
1506          *
1507          * @see FileBackendStore::resolveStoragePath()
1508          *
1509          * @param $storagePath string
1510          * @return Array (container, path) or (null, null) if invalid
1511          */
1512         final protected function resolveStoragePathReal( $storagePath ) {
1513                 list( $container, $relPath, $cShard ) = $this->resolveStoragePath( $storagePath );
1514                 if ( $cShard !== null ) {
1515                         return array( $container, $relPath );
1516                 }
1517                 return array( null, null );
1518         }
1519
1520         /**
1521          * Get the container name shard suffix for a given path.
1522          * Any empty suffix means the container is not sharded.
1523          *
1524          * @param $container string Container name
1525          * @param $relStoragePath string Storage path relative to the container
1526          * @return string|null Returns null if shard could not be determined
1527          */
1528         final protected function getContainerShard( $container, $relPath ) {
1529                 list( $levels, $base, $repeat ) = $this->getContainerHashLevels( $container );
1530                 if ( $levels == 1 || $levels == 2 ) {
1531                         // Hash characters are either base 16 or 36
1532                         $char = ( $base == 36 ) ? '[0-9a-z]' : '[0-9a-f]';
1533                         // Get a regex that represents the shard portion of paths.
1534                         // The concatenation of the captures gives us the shard.
1535                         if ( $levels === 1 ) { // 16 or 36 shards per container
1536                                 $hashDirRegex = '(' . $char . ')';
1537                         } else { // 256 or 1296 shards per container
1538                                 if ( $repeat ) { // verbose hash dir format (e.g. "a/ab/abc")
1539                                         $hashDirRegex = $char . '/(' . $char . '{2})';
1540                                 } else { // short hash dir format (e.g. "a/b/c")
1541                                         $hashDirRegex = '(' . $char . ')/(' . $char . ')';
1542                                 }
1543                         }
1544                         // Allow certain directories to be above the hash dirs so as
1545                         // to work with FileRepo (e.g. "archive/a/ab" or "temp/a/ab").
1546                         // They must be 2+ chars to avoid any hash directory ambiguity.
1547                         $m = array();
1548                         if ( preg_match( "!^(?:[^/]{2,}/)*$hashDirRegex(?:/|$)!", $relPath, $m ) ) {
1549                                 return '.' . implode( '', array_slice( $m, 1 ) );
1550                         }
1551                         return null; // failed to match
1552                 }
1553                 return ''; // no sharding
1554         }
1555
1556         /**
1557          * Get the sharding config for a container.
1558          * If greater than 0, then all file storage paths within
1559          * the container are required to be hashed accordingly.
1560          *
1561          * @param $container string
1562          * @return Array (integer levels, integer base, repeat flag) or (0, 0, false)
1563          */
1564         final protected function getContainerHashLevels( $container ) {
1565                 if ( isset( $this->shardViaHashLevels[$container] ) ) {
1566                         $config = $this->shardViaHashLevels[$container];
1567                         $hashLevels = (int)$config['levels'];
1568                         if ( $hashLevels == 1 || $hashLevels == 2 ) {
1569                                 $hashBase = (int)$config['base'];
1570                                 if ( $hashBase == 16 || $hashBase == 36 ) {
1571                                         return array( $hashLevels, $hashBase, $config['repeat'] );
1572                                 }
1573                         }
1574                 }
1575                 return array( 0, 0, false ); // no sharding
1576         }
1577
1578         /**
1579          * Get a list of full container shard suffixes for a container
1580          *
1581          * @param $container string
1582          * @return Array
1583          */
1584         final protected function getContainerSuffixes( $container ) {
1585                 $shards = array();
1586                 list( $digits, $base ) = $this->getContainerHashLevels( $container );
1587                 if ( $digits > 0 ) {
1588                         $numShards = pow( $base, $digits );
1589                         for ( $index = 0; $index < $numShards; $index++ ) {
1590                                 $shards[] = '.' . wfBaseConvert( $index, 10, $base, $digits );
1591                         }
1592                 }
1593                 return $shards;
1594         }
1595
1596         /**
1597          * Get the full container name, including the wiki ID prefix
1598          *
1599          * @param $container string
1600          * @return string
1601          */
1602         final protected function fullContainerName( $container ) {
1603                 if ( $this->wikiId != '' ) {
1604                         return "{$this->wikiId}-$container";
1605                 } else {
1606                         return $container;
1607                 }
1608         }
1609
1610         /**
1611          * Resolve a container name, checking if it's allowed by the backend.
1612          * This is intended for internal use, such as encoding illegal chars.
1613          * Subclasses can override this to be more restrictive.
1614          *
1615          * @param $container string
1616          * @return string|null
1617          */
1618         protected function resolveContainerName( $container ) {
1619                 return $container;
1620         }
1621
1622         /**
1623          * Resolve a relative storage path, checking if it's allowed by the backend.
1624          * This is intended for internal use, such as encoding illegal chars or perhaps
1625          * getting absolute paths (e.g. FS based backends). Note that the relative path
1626          * may be the empty string (e.g. the path is simply to the container).
1627          *
1628          * @param $container string Container name
1629          * @param $relStoragePath string Storage path relative to the container
1630          * @return string|null Path or null if not valid
1631          */
1632         protected function resolveContainerPath( $container, $relStoragePath ) {
1633                 return $relStoragePath;
1634         }
1635 }
1636
1637 /**
1638  * FileBackendStore helper function to handle file listings that span container shards.
1639  * Do not use this class from places outside of FileBackendStore.
1640  *
1641  * @ingroup FileBackend
1642  */
1643 class FileBackendStoreShardListIterator implements Iterator {
1644         /* @var FileBackendStore */
1645         protected $backend;
1646         /* @var Array */
1647         protected $params;
1648         /* @var Array */
1649         protected $shardSuffixes;
1650         protected $container; // string
1651         protected $directory; // string
1652
1653         /* @var Traversable */
1654         protected $iter;
1655         protected $curShard = 0; // integer
1656         protected $pos = 0; // integer
1657
1658         /**
1659          * @param $backend FileBackendStore
1660          * @param $container string Full storage container name
1661          * @param $dir string Storage directory relative to container
1662          * @param $suffixes Array List of container shard suffixes
1663          * @param $params Array
1664          */
1665         public function __construct(
1666                 FileBackendStore $backend, $container, $dir, array $suffixes, array $params
1667         ) {
1668                 $this->backend = $backend;
1669                 $this->container = $container;
1670                 $this->directory = $dir;
1671                 $this->shardSuffixes = $suffixes;
1672                 $this->params = $params;
1673         }
1674
1675         public function current() {
1676                 if ( is_array( $this->iter ) ) {
1677                         return current( $this->iter );
1678                 } else {
1679                         return $this->iter->current();
1680                 }
1681         }
1682
1683         public function key() {
1684                 return $this->pos;
1685         }
1686
1687         public function next() {
1688                 ++$this->pos;
1689                 if ( is_array( $this->iter ) ) {
1690                         next( $this->iter );
1691                 } else {
1692                         $this->iter->next();
1693                 }
1694                 // Find the next non-empty shard if no elements are left
1695                 $this->nextShardIteratorIfNotValid();
1696         }
1697
1698         /**
1699          * If the iterator for this container shard is out of items,
1700          * then move on to the next container that has items.
1701          * If there are none, then it advances to the last container.
1702          */
1703         protected function nextShardIteratorIfNotValid() {
1704                 while ( !$this->valid() ) {
1705                         if ( ++$this->curShard >= count( $this->shardSuffixes ) ) {
1706                                 break; // no more container shards
1707                         }
1708                         $this->setIteratorFromCurrentShard();
1709                 }
1710         }
1711
1712         protected function setIteratorFromCurrentShard() {
1713                 $suffix = $this->shardSuffixes[$this->curShard];
1714                 $this->iter = $this->backend->getFileListInternal(
1715                         "{$this->container}{$suffix}", $this->directory, $this->params );
1716         }
1717
1718         public function rewind() {
1719                 $this->pos = 0;
1720                 $this->curShard = 0;
1721                 $this->setIteratorFromCurrentShard();
1722                 // Find the next non-empty shard if this one has no elements
1723                 $this->nextShardIteratorIfNotValid();
1724         }
1725
1726         public function valid() {
1727                 if ( $this->iter == null ) {
1728                         return false; // some failure?
1729                 } elseif ( is_array( $this->iter ) ) {
1730                         return ( current( $this->iter ) !== false ); // no paths can have this value
1731                 } else {
1732                         return $this->iter->valid();
1733                 }
1734         }
1735 }